It supports the GSS advanced statistical compact model for extraction and parameter generation methodologies, and can use the power of cluster computing technology to perform statistical simulation and verification tasks rapidly on unprecedented statistical scale.
RandomSpice provides designers with the ability to perform power–performance–yield (PPY) analysis in the presence of acute statistical variability in advanced CMOS technology generations.
Using information obtained from GARAND (the GSS statistical device simulation software) and captured by the statistical compact modeling tool MYSTIC, accurate information about statistical variability can be used with confidence from the early stages of PDK development.
The advanced statistical compact model parameter generation techniques used in RandomSpice allow the accurate reproduction of the shapes and tails of non-Normally distributed extracted statistical compact model parameters and their correlations.
Capturing and maintaining non-Normal behaviour in the compact model parameter generation process, provides significant advantages in SRAM yield analysis.
Statistical enhancement techniques embedded in RandomSpice dramatically reduce the number of Monte Carlo simulations in the design process. Integration with GSS’s ‘push button’ cluster technology lets RandomSpice take advantage of massively parallel task-farming in order to reduce time taken to statistically characterise circuits and standard cells.
According to Professor Asenov, CEO of GSS, “Statistical variability introduced by the discreteness of charge and granularity of matter in present and future CMOS technology is a big challenge but also a big opportunity for the semiconductor industry.
"Understanding and tackling statistical variability can provide significant competitive advantages and can differentiate the players in the semiconductor market”